I live in a small condo with ~40 mains-powered Z-Wave+ devices. The mesh is strong and everything hums about happily
My primary use case is for lighting control, and I use a half-dozen master Scenes to set all-devices just the way I like them for a given activity. They usually fire quite well, however I recognize that a lot of data is being sent very quickly through the mesh, and it's not uncommon for a few devices to hang every once in a while until the storm calms down.
I've considered replacing my many Inovelli dimmers with a Caséta system... but that'll be pricy as heck and rather irresponsible of me
Alternatively, do you think it would be effective to purchase a second C7, and use Hub Link to split my Z-Wave mesh in 2? The intention would be to reduce the size of the data storm when firing scenes that target many devices. Curious to hear your thoughts!
You could also group the devices you use in a scene, and use group metering to stagger commands sent to devices in a group. Maybe something like 30-50ms between devices.
Hello @aaiyar! I've never looked into scene metering. A brief forum search shows that could be a very compelling option. Thanks for opening my eyes to it -- I'll check it out and report back!
I, too, have an entire house full of Inovelli switches. I love the fancy little light bar on them. I primarily use them to indicate the current lighting mode (motion lighting, motion lighting disabled in just this room, motion lighting disabled housewide, etc). I found that when I would disable whole-house motion lighting and my app would need to send a color & level to every strip that it would effectively wedge the egress queue on the Hubitat's Zwave radio. I could actually still receive Zwave messages, just not send.
Anyway, once I figured out the trigger was sending a lot of Zwave commands in a short period of time I whipped up a rudimentary queueing system in my app to put a 100ms second gap between any Zwave commands in my "queue".
Long story short, I think our root problems are the same. Since (it appears) scene metering simply implements a command queueing system like I did, I agree with @aaiyar -- that should get ya' fixed up.
Would you mind sharing that bit of code? I use an app to update the LEDs and it has issues with this... I’m thinking I could add this to my copy of the code to see if it helps.
I can confirm that playing with the Metering setting has greatly improved the success rate of all devices responding properly to a Scene activation... nothing misses or gets left behind The hub still slows down for ~a minute after a Scene is activated, but I can live with that for now!
